Skip to content

Latest commit

 

History

History
391 lines (212 loc) · 9.14 KB

Edip Advanced Configurable Filters.rst

File metadata and controls

391 lines (212 loc) · 9.14 KB
Raw

Advanced Configurable Filters

Kernel filters

Sharpen kernel filter

You can apply a fully configurable :ref:`Sharpen kernel <sharpen-matrix>` filter to the image.

Note

You can preview the kernel you have configurated by using the :guilabel:`&Preview kernel` button.

Emboss kernel filter

You can apply a fully configurable :ref:`Emboss kernel <emboss-matrix>` filter to the image.

Note

You can preview the kernel you have configurated by using the :guilabel:`&Preview kernel` button.

Mean kernel filter

You can apply a fully configurable :ref:`Mean kernel <mean-matrix>` filter to the image.

Note

You can preview the kernel you have configurated by using the :guilabel:`&Preview kernel` button.

Gaussian kernel filter

You can apply a fully configurable :ref:`Gaussian kernel <gaussian-matrix>` filter to the image.

Note

You can preview the kernel you have configurated by using the :guilabel:`&Preview kernel` button.

Kirsch kernel filter

You can apply a fully configurable :ref:`Kirsch kernel <kirsch-matrix>` filter to the image.

Note

You can preview the kernel you have configurated by using the :guilabel:`&Preview kernel` button.

Photography filters

Pencil Sketch

The Pencil Sketch filter produce a non-photorealistic line drawing image.

  • Configuration options:

    • Sigma S:

      Value between 0.0 and 200.0.

    • Sigma R:

      Value between 0.0 and 1.0.

    • Shade factor:

      Value between 0.001 and 0.100.

Stylisation

The stylisation filter is an edge-aware filters which effect is not focused on photorealism but

abstract regions of low contrast while preserving, or enhancing, high-contrast features.

  • Configuration options:

    • Sigma S:

      Value between 0.0 and 200.0.

    • Sigma R:

      Value between 0.0 and 1.0.

Detail Enhance

The Detail Enhance filter enhances the details of a particular image.

  • Configuration options:

    • Sigma S:

      Value between 0.0 and 200.0.

    • Sigma R:

      Value between 0.0 and 1.0.

Edge Preserving

The Edge Preserving filter is a smoothing filters used in many different applications.

  • Configuration options:

    • Sigma S:

      Value between 0.0 and 200.0.

    • Sigma R:

      Value between 0.0 and 1.0.

    • Filter:

      • Recurse filter.
      • NormConv Filter.

Denoising

The denoising filter perform image denoising using Non-local Means Denoising with several computational optimizations.

Noise expected to be a gaussian white noise.

Denoising is done by converting image to CIELAB colorspace and then separately denoise L and AB components with different parameters.

  • Configuration options:

    • Luminance factor:

      Parameter regulating filter strength. Big values perfectly removes noise but also removes image details, smaller values preserves details but also preserves some noise.

    • Color denoising factor:

      The same as Luminance factor but for color components.

Morphological filters

Morphological operations process images according to shapes.

They apply a defined structuring element (named :ref:`kernel <convolution-matrix>`) to the image obtaining a new image where the current pixel is computed by comparing it to his neighborhood pixels.

Note

Depending on the structuring element selected (which is fully configurable in Edip)

a morphological operation is more sensitive to one specific shape or the other.

Morphological operations

Dilatation

Dilatation add pixels from the background to the boundaries of the object in an image.

In dilatation, the value of the output pixel is the maximum of all pixels in the neighborhood.

Erosion

Erosion remove pixels from the foreground.

Using erosion the value of the output pixel is the minimum of all pixels in the neighborhood.

Opening

Opening is an :ref:`erosion <morphological-erosion>` followed by a :ref:`dilatation <morphological-dilatation>`.

Opening remove small objects from an image while preserving the larger one.

Closing

Closing is a :ref:`dilatation <morphological-dilatation>` followed by an :ref:`erosion <morphological-erosion>`.

Closing is used to remove samll holes while preserving the larger one.

Tophat

Tophat ouput is the difference between the source image and it's :ref:`opening <morphological-open>`.

Blackhat

Blackhat ouput is the difference between the source image and it's :ref:`closing <morphological-close>`.

Canny filter

The Canny contours detection operator use 2 different :ref:`threshold <thresholding>` values:

Low threshold:

The low threshold should be chosen in a way that it included all edges pixels that are considered as to belong to a significant image contour.

This should detect more edges that what is ideally needed in the case of the Canny algorithm.

High threshold:

His role is to define the edges that belong to all important contours.

It should exclude all edges considered as outlier.

The Canny algorithm combines these two edges maps in order to produce an optimal map of contours.

It operate by keeping only edge points of the low threshold edge map for which a continuous path of edges exist, linking all edges points of the high threshold map are kept, while isolated chains of edges points in the low threshold map are removed.

Note

This strategy, based on the use of two threshold to obtain a binary map, is called histereris thresholding.

In addition, the Canny algorithm uses an extra strategy to improve the quality of the edge map: all edges points for which the gradient magnitude is not a maximum in the gradient direction are removed.

The gradient orientation is always perpendicular to the edge. Therefor, the local maximum of the gradient is this direction correspond to the points of maximum strength of the contours.

This explains why thin edges are obtained in tha Canny contours map.

Color contours filter

This produce the same result as the Canny filter except that you can set:

  • The foreground color: the color of the detected edges.
  • The background color.